Millisecond Forums

Inquisit Web using group id generation with random selection without replacement - counterbalanced?

https://forums.millisecond.com/Topic29578.aspx

By linatet - 8/18/2020

Hi,
There are 12 condition in my iqx, each created like:

 <expt>
/subjects = (12 of 12)
/groupassignment = groupnumber
/ blocks = [
    1=inst;
    2=targetincompatiblepractice;
    3=attributepractice;
    4=incompatibletest1
]
/onexptend = [values.completed = 1]
</expt>

I want participants to be randomly assigned to each experiment condition, and that equal numbers of participants go to each condition. On Inquisit Web, there is the option to choose how the group ID will be determined - random selection without replacement. Will this accomplish what I want? Or, do I need to create batches for that? Also, what are the situations in which someone would use batch to assign instead of expt?
By Dave - 8/18/2020

 
linatet - 8/18/2020
Hi,
There are 12 condition in my iqx, each created like:

 <expt>
/subjects = (12 of 12)
/groupassignment = groupnumber
/ blocks = [
    1=inst;
    2=targetincompatiblepractice;
    3=attributepractice;
    4=incompatibletest1
]
/onexptend = [values.completed = 1]
</expt>

I want participants to be randomly assigned to each experiment condition, and that equal numbers of participants go to each condition. On Inquisit Web, there is the option to choose how the group ID will be determined - random selection without replacement. Will this accomplish what I want? Or, do I need to create batches for that? Also, what are the situations in which someone would use batch to assign instead of expt?

> On Inquisit Web, there is the option to choose how the group ID will be determined - random selection without replacement. Will this accomplish what I want?

Select either sequential generation or random without replacement and indicate the number of groups as 12. You will still get imbalances across conditions -- an ID is generated whenever someone visits (or refreshes) the experiment's start page, regardless of whether that person then actually starts the experiment or not.